#!/usr/bin/env python
# coding: utf-8

# # Iris Data Classification
# - Link Component Color Annotations
#   - <font color='yellow'>Yellow</font> : data load / preprocessing
#   - <font color='green'>Green</font> : EDA
#   - <font color='violet'>Violet</font> : model train / predict

# ### Required Python Packages
# - `numpy`
# - `pandas`
# - `scikit-learn`
# - `seaborn`
# - `matplotlib`
# 
# Run the following cell to install the packages.

# In[ ]:


#
# Required Packages
#   Run this cell to install required packages.
#
get_ipython().run_line_magic('pip', 'install "matplotlib>=2.0" "numpy>=1.19" "pandas>=1.1" "scikit-learn>=0.22.2" "seaborn>=0.11"')


# ### 0. Global Parameters
# - global paprameter of link pipeline
#   - test_size : rate of valid-set when train-valid-split
#   - random_state : random_state 

# ### 1. Load package,data

# In[ ]:


import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
from sklearn.datasets import load_iris
from sklearn.ensemble import RandomForestClassifier, VotingClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score
from sklearn.model_selection import train_test_split
from sklearn.neighbors import KNeighborsClassifier


# In[ ]:


iris = load_iris()


# In[ ]:


df = pd.DataFrame(data=iris.data, columns=iris.feature_names)
df["target"] = iris.target
display(df)


# The `df` of green components and the `df` of purple components can be used independently of each other from the point of branching
# 
# * Modeling -> `target` of `df` : **0,1,2**
# * EDA -> `target` of `df` : **setosa, versicolor, virginica**

# ### 2. EDA

# In[ ]:


df["target"] = df["target"].map({0: "setosa", 1: "versicolor", 2: "virginica"})
display(df)


# In[ ]:


sns.pairplot(df, x_vars=["sepal length (cm)"], y_vars=["sepal width (cm)"], hue="target", height=5)
sns.pairplot(df, x_vars=["petal length (cm)"], y_vars=["petal width (cm)"], hue="target", height=5)
plt.show()


# In[ ]:


plt.figure(figsize=(7, 4))
sns.heatmap(
    df.corr(), annot=True, cmap="coolwarm"
)  # draws  heatmap with input as the correlation matrix calculted by(iris.corr())
plt.show()


# ### 3. Modeling

# In[ ]:


df_X = df.drop("target", axis=1)
df_y = df["target"]
train_X, valid_X, train_y, valid_y = train_test_split(
    df_X, df_y, test_size=test_size, random_state=random_state, stratify=df_y
)


# In[ ]:


model = LogisticRegression(random_state=random_state)
model.fit(train_X, train_y)


# In[ ]:


pred = model.predict(valid_X)
print(f"Accuracy : {accuracy_score(valid_y, pred)}")


# In[ ]: